Date: 2025-08-20 Page is: DBtxt003.php txt00028718 | |||||||||
|
CORPORATE MANAGEMENT
IMPROVING PERFORMANCE TPBopinion: Fixing a mainframe computer system at Aerosol Techniques Inc (ATI) Original article: | |||||||||
|
When I was in my late 20s I was recruited into a fairly responsible job in the US corporate world. I was recruited as an assistant to the Treasurer and Chief Financial Officer of a company called Aerosol Techniques Inc. (ATI) in Milford, Connecticut. My initial task was to help the CFO identify the biggest constraints on profit performance and recommend corrective aaction.
The CFO was quite young and not long in the job. He was a graduate of the Harvard Business School and previously had qualified in the UK as a Chartered Accountant. Though he had a Harvard MBA he did not have any undergraduate degree! He had convinced Harvard that being a British (England and Wales) Chartered Accountant was equivalent or better than any undergraduate degree that was the basis for Harvard Business School entrance!
I had similar accountancy training in the UK, but I also had a Cambridge University undergraduate degree. Better yet, I had studied both engineering and economics at Cambridge before getting some industrial management experience and then working through 'Articles' in Chartered Accoutnancy.with Cooper Brothers & Co. (CB&Co), one of the top accountancy firms in London.
The biggest profit performance issues were pretty obvious. The company had four divisions, plus a corporate office. Three divisions were reasonably profitable. The fourth division which was integrated with the corporate office was not. It was also the newest, and home to a costly computerised 'management system'.
High profile consultants had failed to solve the computer issues and profit performance at this division. The Division Controller was let go, and I was put into the position. This was also the department responsible for the operation of the computer.
Almost immediately it became obvious that the computer was processing inaccurate input data. I remember talking about GIGO ... Garbage In, Garbage Out. But 'why' was the input data wrong.
Almost immdeiately it became obvious that the factory operations and the computer were operating on two very differnt time cycles. Factory production was moving a lot more rapidely than the information about the production and the related inventory movements being recorded by the computer system,
When the computer information was used to make inventory decisions, the inventory in actual reality was different from the computer inventory. The reason was that they were differnt pieces of data. And when they were 'corrected', they were in fact being made 'more wrong'. Very quickly the computer information degraded into a useless pile of wrong numbers. The factory had kept running by a maze of informal manual systems that completely ignored the computor system. And it was, in fact, better than the computer system.
The 'factory' had concluded that the computer was incapable of anything ... anbd really had given up on trying!
I needed to do 'something' to demonstrate that the computer was really good at calculating and printing reports. I set about computerising the 'fixed asset ledger' The production / inventory data and the fixed asset data had a lot in common from an accounting and computational perspective. The big differnce was tha production / inventory changed very quickly and fized asset information changed slowly.
Everyone was very impressed when I printed out the new computerised fixed asset information. When they checked it, they found that it was accurate with all the depeciation computed accurately. Of course, it was a million times easier to do than the fast moving computerised production / inventory control system ... but the core logic was the same.
It was not obvious haw to get control of the dataflows from raw material purchasing into inventory through production into finished goods inentory and finally shipping / cales. All these components existed in the system but the data being processed was not 100% correct. While some of the inventory moved very slowly, maybe 90% of the inventory came and went in less than a day.
Which begs the question ... what is a day? Most of the time, this factory ran a two shift operation, but from time to time an addional shirt was added to handle sn urgent surge in production. Getting the computer and the factory to be 'in sync' was absolutley essential for the computerised production system to function correctly.
I made some progress by insisting that all 'adjustments' were made at very specific times of the day ... at a time when the correction in the warehouse would happen at the same time as the computer record was updated. The logic of this was OK, but in practice it did not work very well. The data entry key-punch group was always behind, but not always by the same amount. We got some improvement in data reality/accuracy, but by no means enough.
At this point we were still operating a vomputerised 'locator' system in the warehouse for both raw materials and finished goods. Every pallet of raw materials or of finished goods was identified and its movements and location in the warehouse put into a computerised record. It was expected that every movement in the warehouse would be recorded in the locator system. This was a huge amount of (computerised) paperwork for rather little incremental information. It took me a while, but eventually I 'pulled the plug' and closed down this computerised 'locator system'.
co
One of my tasks was to get a computer to work that my employer had acquired from IBM. After almost two years of getting advice from IBM it still did nothing useful for the company. I became the 'controller' of this division of the company that had the computer and responsible for its operation.
Peter Burgess COMMENTARY Peter Burgess | |||||||||